-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathActiveRAdam.py
169 lines (136 loc) · 7.53 KB
/
ActiveRAdam.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
import math
import torch
from torch.optim.optimizer import Optimizer, required
class ActiveRAdam(Optimizer):
r"""Implements AdamW algorithm.
The original Adam algorithm was proposed in `Adam: A Method for Stochastic Optimization`_.
The AdamW variant was proposed in `Decoupled Weight Decay Regularization`_.
Arguments:
params (iterable): iterable of parameters to optimize or dicts defining
parameter groups
lr (float, optional): learning rate (default: 1e-3)
betas (Tuple[float, float], optional): coefficients used for computing
running averages of gradient and its square (default: (0.9, 0.999))
eps (float, optional): term added to the denominator to improve
numerical stability (default: 1e-8)
weight_decay (float, optional): weight decay coefficient (default: 1e-2)
amsgrad (boolean, optional): whether to use the AMSGrad variant of this
algorithm from the paper `On the Convergence of Adam and Beyond`_
(default: False)
.. _Adam\: A Method for Stochastic Optimization:
https://arxiv.org/abs/1412.6980
.. _Decoupled Weight Decay Regularization:
https://arxiv.org/abs/1711.05101
.. _On the Convergence of Adam and Beyond:
https://openreview.net/forum?id=ryQu7f-RZ
"""
def __init__(self, params, stepSize, lr=1e-3, betas=(0.9, 0.999), eps=1e-8,
weight_decay=1e-2, amsgrad=False, lrHigh=.05, lrLow=.95):
if not 0.0 <= lr:
raise ValueError("Invalid learning rate: {}".format(lr))
if not 0.0 <= eps:
raise ValueError("Invalid epsilon value: {}".format(eps))
if not 0.0 <= betas[0] < 1.0:
raise ValueError("Invalid beta parameter at index 0: {}".format(betas[0]))
if not 0.0 <= betas[1] < 1.0:
raise ValueError("Invalid beta parameter at index 1: {}".format(betas[1]))
if not 0.0 <= weight_decay:
raise ValueError("Invalid weight_decay value: {}".format(weight_decay))
defaults = dict(lr=lr, betas=betas, eps=eps,
weight_decay=weight_decay, amsgrad=amsgrad,
lrHigh=lrHigh, lrLow=lrLow, stepSize=stepSize)
self.buffer = [[None, None, None] for ind in range(10)]
super(RAdamConsciousLR, self).__init__(params, defaults)
def __setstate__(self, state):
super(ActiveRAdam, self).__setstate__(state)
for group in self.param_groups:
group.setdefault('amsgrad', False)
@torch.no_grad()
def step(self, closure=None):
"""Performs a single optimization step.
Arguments:
closure (callable, optional): A closure that reevaluates the model
and returns the loss.
"""
loss = None
if closure is not None:
with torch.enable_grad():
loss = closure()
for group in self.param_groups:
for p in group['params']:
if p.grad is None:
continue
# Perform optimization step
grad = p.grad
if grad.is_sparse:
raise RuntimeError('Adam does not support sparse gradients, please consider SparseAdam instead')
amsgrad = group['amsgrad']
state = self.state[p]
# State initialization
if len(state) == 0:
state['step'] = 0
# Exponential moving average of gradient values
state['exp_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
# Exponential moving average of squared gradient values
state['exp_avg_sq'] = torch.zeros_like(p, memory_format=torch.preserve_format)
if amsgrad:
# Maintains max of all exp. moving avg. of sq. grad. values
state['max_exp_avg_sq'] = torch.zeros_like(p, memory_format=torch.preserve_format)
state['gai'] = torch.ones_like(p, memory_format=torch.preserve_format)
state['cumm'] = torch.zeros_like(p, memory_format=torch.preserve_format)
# Accumulate gradients for the epoch
state['cumm']+=(p.grad)
# print('step', state['step'], 'cumm', state['cumm'], 'grad', p.grad.item())
exp_avg, exp_avg_sq = state['exp_avg'], state['exp_avg_sq']
# print('exp_avg',state['exp_avg'])
if amsgrad:
max_exp_avg_sq = state['max_exp_avg_sq']
beta1, beta2 = group['betas']
state['step'] += 1
bias_correction1 = 1 - beta1 ** state['step']
bias_correction2 = 1 - beta2 ** state['step']
# Decay the first and second moment running average coefficient
# exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1)
state['exp_avg'] = beta1 * (exp_avg) + (1-beta1)*(grad)
# print('exp_avg',state['exp_avg'])
# exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2)
state['exp_avg_sq'] = beta2 * exp_avg_sq + (1-beta2)*grad.pow(2)
# RAdam Start1
buffered = self.buffer[int(state['step'] % 10)]
if state['step'] == buffered[0]:
N_sma, step_size = buffered[1], buffered[2]
else:
buffered[0] = state['step']
beta2_t = beta2 ** state['step']
N_sma_max = 2 / (1 - beta2) - 1
N_sma = N_sma_max - 2 * state['step'] * beta2_t / (1 - beta2_t)
buffered[1] = N_sma
# more conservative since it's an approximated value
if N_sma >= 5:
step_size = group['lr'] * math.sqrt((1 - beta2_t) * (N_sma - 4) / (N_sma_max - 4) * (N_sma - 2) / N_sma * N_sma_max / (N_sma_max - 2)) / (1 - beta1 ** state['step'])
else:
step_size = group['lr'] / (1 - beta1 ** state['step'])
buffered[2] = step_size
# RAdam End1
# Perform stepweight decay
p.mul_(1 - group['lr'] * state['gai'] * group['weight_decay'])
# more conservative since it's an approximated value
if N_sma >= 5:
denom = exp_avg_sq.sqrt().add_(group['eps'])
# p_data_fp32.addcdiv_(-step_size, exp_avg, denom)
p -= step_size*state['gai']*(exp_avg/denom)
else:
# p_data_fp32.add_(-step_size, exp_avg)
p -= step_size*state['gai']*exp_avg
# SetLR if i>0
if state['step']/group['stepSize'] > 1 and state['step']%group['stepSize']==0:
tmp2 = state['gradOld'].clone().cpu()##could be eliminated
tmp3 = state['cumm'].clone().cpu()##could be eliminated
tmp5 = state['gai'].clone().cpu()##may be the one that needs cloning
state['gai'] = torch.as_tensor(np.where(tmp2*tmp3<=0, tmp5.mul(group['lrLow']), tmp5.add(group['lrHigh'])),dtype=p.dtype , device=p.device)
# Resetting the accumulated gradients after each epoch
if state['step']%group['stepSize']==0:
cumm = state['cumm']
state['gradOld'] = cumm.clone()
state['cumm'] = torch.zeros_like(p, memory_format=torch.preserve_format)
return loss